First off, I believe that the FCS data standard has done an outstanding job of
allowing (1) the exchange of primary data; and (2) the generation of some
excellent third-party software to analyze flow cytometry data. Both of these
are beneficial to our community at large! These goals must be kept in mind in
the evolution of the standard.
Therefore, I strongly support that continued use of FCS as a data standard for
flow cytometry. I think many of the modifications in FCS3.0 are good, and hope
that the manufacturers who write acquisition software will adopt FCS3.0 as soon
as it has been "ratified" (or whatever process is required for official ISAC
blessing).
The principle problem with FCS, from my standpoint, is that the specification of
data types is far too broad. This is a potential problem, because no one has
written a generally-distributed software that will read in all the different
formats allowed! Why? Because no one uses a vast majority of these formats?
Why should I incorporate the ability to read IEEE floating point numbers for
parameter values in my software when this feature would be used by perhaps one
laboratory in the world? The cost-benefit ratio does not make such coding
viable.
It is readily apparent that a vast majority of FCS files generated at this time
use a single format of the data. Therefore, I have proposed a "simplified" FCS
standard (copies of this proposal have been sent to G. Valet and D. Coder for
distribution on the appropriate web sites; I will send it by EMail to anyone who
requests it as well). "FCS3.0S" is designed to keep the vast majority of extant
data files compliant, and yet assure that future software development efforts
will yield applications that can read all possible compliant files.
The draft of this proposal is just that: a draft. I didn't spend all that much
time on it, because I don't know how receptive the FCS committee will be (or the
community at large). If there IS interest, please send comments and criticisms
to me (or to the discussion sites).
mr